mN

l a, P

y Med

of Lav

Veterinary Microbiology 144 (2010) 371376

Article history:

Received 22 August 2009

Received in revised form 15 January 2010

The goals of this studywere to develop a PCR technique to detect ascV and aopB genes from

the type III secretion system (T3SS), to evaluate the frequency of these genes in Aeromonas

hydrophila strains isolated from diseased sh and from aquaculture environments, and to

Contents lists available at ScienceDirect

Veterinary Mi

journa l homepage: www.e ls1. Introduction

Aeromonas hydrophila, an ubiquitous bacterium inaquatic environments, has been isolated from drinkingwater and food, thus it is a serious concern to public andanimal health (Edberg et al., 2007). It has been associatedwith illness in a broad spectrum of hosts as reptiles,amphibians, mammals (Vivas et al., 2004) and sh, causingoutbreaks in sh farms with high mortality rates andelevated economic losses to the aquaculture industry(Fang et al., 2004).

The pathogenesis of A. hydrophila infections is complexinvolving several virulence factors. Among these the type

III secretion system (T3SS), has been considered to be ofgreat importance (Chacon et al., 2004; Vilches et al., 2004;Yu et al., 2004; Sha et al., 2005; Cornelis, 2006; Sierra et al.,2007; Mueller et al., 2008). The genes ascV and aopB,essential components of the A. hydrophila T3SS, codifyproteins associated with the xation of the T3SS in thebacterial inner membrane and with assembly of thetranslocon, respectively. Because of these functions, ascVand aopB genes are reasonable constituents of the T3SS tobe screened in bacterial strains. Chacon et al. (2004)reported no signicant differences between the presenceof T3SS genes in A. veronii, A. caviae, and A. hydrophilarecovered from human cases of intestinal and extraintest-inal illness. Likewise, Vilches et al. (2004) demonstratedthat the frequency of T3SS genes was higher in isolatesfrom human clinical cases when compared with environ-mental strains. However, no data exist about the com-parative frequency of these genes in A. hydrophila isolatedfrom diseased sh and from aquaculture environments.

Accepted 25 January 2010

Keywords:

Aeromonas hydrophila

Type III secretion system

Diseased sh

Aquaculture environment

determine the relationship between the presence of these genes and virulence of A.

hydrophila in Nile tilapia. The PCR assay developed here successfully detected the target

genes, showing three different proles for the strains ascV+/aopB+, ascV+/aopB, andascV/aopB. A higher frequency of ascV+/aopB+was veried in isolates fromdiseased shcompared to those from aquaculture environments (P< 0.05). Among 64 isolates from

diseased sh, ascV+/aopB+ (62.5%) was the most frequent prole (P< 0.05) and caused

more intensive mortality rates. Environmental strains containing the ascV+/aopB+ prole

were less virulent than isolates from clinical cases. These results suggest that the presence

of a functional T3SS probably increases the virulence of A. hydrophila. The PCR technique

was shown to be a specic and efcient tool for detection of T3SS, and this technique can

be used for virulence typing of A. hydrophila isolates.

2010 Elsevier B.V. All rights reserved.

* Corresponding author at: Federal University of Lavras, Veterinary

Medicine, UFLAs Campus, 37200-000 Lavras, Minas Gerais, Brazil.

Tel.: +55 35 3829 1714; fax: +55 35 3829 1715.

E-mail address: henrique@ua.br (H.C.P. Figueiredo).

0378-1135/$ see front matter 2010 Elsevier B.V. All rights reserved.Detection of type III secretion systetheir relationship with virulence in

G.A. Carvalho-Castro a, C.O. Lopes a, C.A.G. LeaaAQUAVET, Laboratory of Aquatic Animal Diseases, Department of Veterinarb Laboratory of Molecular Biology, Department of Biology, Federal University

A R T I C L E I N F O A B S T R A C Tdoi:10.1016/j.vetmic.2010.01.021genes in Aeromonas hydrophila andile tilapia

.G. Cardoso b, R.C. Leite a, H.C.P. Figueiredo a,*

icine, Federal University of Lavras, Lavras MG 37200-000, Brazil

ras, Lavras MG 37200-000, Brazil

crobiology

evier .com/ locate /vetmic

buffer (0.04M Trisacetate, pH 8.4, 1mM EDTA) and werevisualized with a UV transilluminator after staining withethidium bromide solution (0.5mg/mL). A 100 bp ladderDNA molecular marker (New England Biolabs, USA) wasused in each electrophoresis assay.

2.5. DNA sequencing

PCRproducts forascVand aopBgeneswerepuriedusinga Wizard PCR Preps kit (Promega) and then sequenced.Sequencing reactions were performed using a BigDyeTM

Terminator Cycle sequencing kit (Applied Biosystems, USA)

G.A. Carvalho-Castro et al. / Veterinary Microbiology 144 (2010) 371376372The goals of this study were to develop PCR protocols todetect ascV and aopB genes in A. hydrophila strains, toevaluate their frequency in isolates from diseased sh andfrom aquaculture environments, as well as to address thein vivo virulence in Nile tilapia (Oreochromis niloticus) ofstrains belonging to the different genetic proles.

2. Materials and methods

2.1. Bacterial strains

From a total of 418 Aeromonas spp. isolates ofAQUAVET bacterial culture collection, 104 strains identi-ed as A. hydrophila were evaluated in this study. Theisolates were obtained from 12 sh farms located in threedifferent Brazilian states (Supplementary Table 1). 64strains were isolated from the brain (n = 6), spleen (n = 5),kidney (n = 33), wounds (n = 14), liver (n = 5), and ascitesuids (n = 1) of diseased sh of the following ve species:O. niloticus, Rhamdia quelen, Piaractus mesopotamicus, Betasplendens, and Brycon orbignyanus. In addition, 40 isolateswere from aquaculture environments (28 from supplywater and 12 from pond water). A. hydrophila isolateswere biochemically characterized following Abbott et al.(2003), and bacterial species conrmed by genotypicidentication using PCR-RFLP analysis of the 16S rRNAgene digested with the enzymes AluI and MboI accordingto Borrel et al. (1997), with minor modication. Briey,the digestion products were resolved using 8% polyacry-lamide gel electrophoresis. The isolates were stored at70 8C until use.

2.2. DNA extraction

The strains were thawed, streaked onto trypticase soyagar (TSA) and incubated at 30 8C for 18 h. A single colonyfrom the culture was resuspended in 50ml of sterileultrapure water, vortexed at high speed for 1min andincubated at 95 8C for 10min in thermocycler for cell lyses.The extracted DNAwas used immediately as a template forPCR reactions. Sterilized ultrapure water was used as anegative control of extraction.

2.3. Oligonucleotide primers

For development of our PCR technique, primer sets todetect the ascV and aopB genes from A. hydrophila strainswere designed. Nucleotide sequences of these genes wereavailable in NCBI database and were obtained and alignedfor the primer design. The construction and preliminaryevaluation of the primers were performed using thesoftwares DNAMAM version 4.0 (Lynnon Corporation,Canada) and BLAST (http://www.ncbi.nlm.nih.gov/BLAST)for characteristics and specicity analysis, respectively.The primers ascV sense and ascV antisense were designedbased on the highly conserved regions of ascV sequencesfrom A. hydrophila strains AH1 and AH3 (GenBankaccession numbers AY394563 and AY528667, respec-tively) (Vilches et al., 2004; Yu et al., 2004) to amplifyan ascV fragment of approximately 891 bp. The primersaopB sense and aopB antisense were constructed based onthe aopB sequence from A. hydrophila strain AH1, with theaim of amplifying an aopB fragment of 951 bp. The fourprimers were synthesized by Integrated DNA Technologies(USA). Table 1 shows the oligonucleotide primers.

2.4. PCR amplication

PCR reactions were standardized using different con-centrations ofMgCl2, 100mMdeoxyribonucleoside tripho-sphate (dNTPs) mixture solution (Amresco, USA), primers,Taq DNA polymerase, and annealing temperature. Theabsence of unspecic products and band intensity of PCRproducts were evaluated for choice of the best reaction.The PCR was performed in a thermocycler using the GoTaqFlexi DNA Polymerase kit (Promega, USA), which included5 buffer and 25mM MgCl2. After standardization, thebest PCR mix for ascV amplication consisted of 1 buffer,0.25mM of each primer, 1.5mM of MgCl2, 0.1mM of eachdNTP (dATP, dTTP, dGTP, and dCTP), 1.25 U of Taq DNApolymerase, and 2ml of DNA template. For the aopB gene,the best amplication was achieved using 1 buffer, 0.25mM of each primer, 3.0mM MgCl2, 0.2mM of each ofdNTP, 3.0 U of Taq DNA polymerase, and 2ml of DNAtemplate. All reactions were completed using sterilizedultrapure water for a total volume of 25ml. A. hydrophilaATCC 7966Twas included as a negative control in each test,since the absence of T3SS in this reference strain veriedafter its whole genome sequencing (Seshadri et al., 2006).The PCR for all isolates was performed twice to conrmreproducibility.

PCR conditions for both genes consisted of an initialdenaturation step at 95 8C for 5min followed by 35 cycles ofamplication inwhichdenaturation, annealing, and elonga-tion temperatures were 95 8C for 1min, 58 8C for 90 s, and72 8C for 1min. The nal elongation was at 72 8C for 5min.

The amplication products were analyzed using 1.5%(w/v) agarose gel electrophoresis with 1 Trisacetate

Table 1

Primers for PCR amplication of genes ascV and aopB from A. hydrophila

strains.

Primer Position in genes

of A. hydrophila

AH1a

Sequence (5030)

ascV sense 509527 AGCAGATGAGTATCGACGG

ascV antisense 13801399 AGGCAT TCTCCTGTACCAG

aopB sense 1633 TACCTGTTGGAATGATTCCG

aopB antisense 947966 AGTGAACGCCCTCTCTCCa Accession number Genbank AY394563.

and 951 bp fragments that resulted from the amplicationof the respective genes. Negative reactions were veriedfor the A. hydrophila ATCC 7966T DNA template andultrapure water, as expected. The sequenced ampliconswere analyzed by the BLASTn algorithm, and a high degreeof similaritywas found between the PCR products obtainedin this work and ascV and aopB gene sequences available in

Fig. 1. The pairs 1, 2 and 3 demonstrate the three genetic proles ascV+/

aopB+, ascV+/aopB and ascV/aopB, respectively, from PCR products ofAeromonas hydrophila strains. Bands of 891 and 951 bp represent,

respectively, ascV and aopB genes.

G.A. Carvalho-Castro et al. / Veterinary Microbiology 144 (2010) 371376 373and run on an ABI 3730XL genetic analyzer (AppliedBiosystems). Sequences were aligned and then comparedwithsequencesofascVgenesandaopBgenes available in theNCBI database using the BLASTn algorithm.

2.6. Experimental infection

Toaddress thevirulenceof isolateswithdifferentgeneticpatterns, 22 A. hydrophila strains were randomly selectedbased on the three proles resulted from PCR reactions.From those, nine were ascV+/aopB+ (positive for the twogenes), nine were ascV/aopB (negative for both genes),and fourwere ascV+/aopB (positive only for the ascV gene).The selected strains were inoculated in tryptic soy brothunder calcium-depleted conditions and incubated at 30 8Cfor 6 h under low agitation (150 rpm). The bacterialsuspension was then adjusted to an optical density of0.180 0.020 at 600 nm, corresponding approximately to106 CFU/mL. Suspensions were serially diluted in 0.5Mphosphate-buffered saline (PBS), streaked onto TSA, andincubated at 30 8C for 18 h for bacterial counting. To preparethe bacterial inoculum, after growth the cells were harvestedby centrifugation (3000 g, 30min), washed once andresuspended in sterile PBS. Nile tilapia (O. niloticus) ngerlingswith an average weight of 25.6 5.3 g were acquired from acommercial hatchery for A. hydrophila challenge and acclima-tized to laboratory conditions by 15 days. Each experimentalgroup consisted of eight sh kept in a 57-L aquarium suppliedwithow-throughdechlorinated tapwater (0.5 L/h). Fishweremaintained on a 12 h:12 h light/dark period at a watertemperature of 26 8C and were fed with VITAFISH 32% PB(Matsuda, Brazil) to satiation four times a day. Fish wereanesthetized by immersion in a bucket containing 100mg/Lbenzocaine. The challenged groups were injected intraper-itoneallywith PBSwashed A. hydrophila cells corresponding to106 CFU/sh and the control sh with 0.1mL of sterile PBS.Challenged sh were monitored for 15 days. Samples of brain,liver, and kidney were aseptically collected from all dead shand inoculated on sheep blood agar to recover the bacteria.After a period of 15 days, all sh (challenged and controlsgroups) were killed by benzocaine overdose and submitted tobacteriologic assay. All in vivo experiments were carried outaccording to animal welfare standards and were approved bythe Ethical Committee for Animal Experiments of the FederalUniversity of Lavras, Brazil.

2.7. Statistical analysis

Fishers Exact Test using SAS1 statistical software STATVersion 6.12 (SAS Institute Inc., USA) was applied todetermine whether the gene frequencies obtained inisolates from different origins were statistically different.A P value of 0.05 or less was considered statisticallysignicant.

3. Results

3.1. PCR standardization

Positive PCR reactions were obtained for the ascV andaopB genes of A. hydrophila isolates. Fig. 1 presents the 891

from

F

1

0

3

4

was s

oles

G.A. Carvalho-Castro et al. / Veterinary Microbiology 144 (2010) 371376374the NCBI database for A. hydrophila (accession numbersAY528667/AY394563 and AY394563, respectively); 98%and 96% identity conrmed the specicity of the primersused in this study. The sequenced products for genes ascVand aopB of strains AE288 and AE110were deposited in theNCBI GenBank database with the accession numbersGU384670, GU384671, GU384672, and GU384673.

3.2. Frequency of ascV and aopB genes

The T3SS genes evaluated in this work were detected inA. hydrophila isolates from all farms and all sh speciestested (Supplementary Table 1). Table 2 shows thedistribution of positive and negative strains according tosh farm and source. Three genetic types occurred in thestrains analyzed: ascV+/aopB+, ascV/aopB, and ascV+/aopB. A total of 53 strains were ascV+/aopB+, of which 40(75.5%) were isolated from diseased sh and 13 (24.5%)from the environment. A signicant difference (P< 0.05)was observed among these sources. Four isolates fromdiseased sh showedwere ascV+/aopB (6.25%). Among 64isolates from diseased sh, ascV+/aopB+ (62.5%) was themost frequently observed prole (P< 0.05).

3.3. Experimental infection

During the experimental infection period, no signs of

Table 2

Frequency of T3SS genes in A. hydrophila isolated from diseased sh and

Strain source Farm

A B C D E

Clinical case

ascV+/aopB+a 2 16 3

ascV+/aopB 0 2 2ascV/aopB 0 1 11

Environment

ascV+/aopB+ 2 6

ascV+/aopB 0 0ascV/aopB 6 12

Total 8 02 19 16 18a The frequency of the prole ascV+/aopB+ in strains from diseased sh

(P< 0.05). No signicant differences were obtained between the other prdisease or mortalities were observed in the control group.Likewise, negative results were veried in bacteriologicalanalyses at the end of experiment, showing absence of A.hydrophila infection in those sh. In the groups challengedwith ascV+/aopB+ strains, the rst signs of disease wereapparent 4 h post-infection. The majority of mortalitiesoccurred by 8 h post-challenge, except for some shwithinthe groups infected with strains AE 110, AE 288, and AE111. For these groups, mortalities occurred from 48 h to 7days. Bacterial reisolation was conducted from brain, liver,and kidney of all dead sh. Themain clinical signs observedin sh groups challenged with ascV+/aopB+ strains werecutaneous hemorrhage at the base of all ns and in themouth, ascites with a serobloody uid, and, in some sh,exophthalmia and erratic swimming. In contrast, in thesh infected with AE 204 (ascV+/aopB) and with AE 190,AE 220, and AE 266 (all belonging to ascV/aopB), themortalities began after 24 h post-inoculation. The diseasedsh presented only anorexia, lethargy, and skin darkening,followed by death.

Mortality rates varied between 12.5% and 87.5% in thegroups challenged with A. hydrophila ascV+/aopB+. StrainsAE 067, AE 110, AE 111, and AE 406 were highly virulent toNile tilapia ngerlings, inducing 75%, 75%, 87.5%, and 87.5%mortality, respectively (Table 3). Strains AE 190, AE 220,and AE 266 (all ascV/aopB) exhibited low virulence inNile tilapia ngerlings, resulting inmortality rates of 37.5%,12.5%, and 12.5%, respectively. Likewise, among the ascV+/aopB strains, only AE 204 caused disease, with amortalityrate of 12.5%. Environmental strains with the ascV+/aopB+prole were less virulent than isolates from clinical cases.

4. Discussion

We developed a PCR method to detect the genes ascVand aopB, which both encode essential components of theT3SS nanomachine. Positive results were obtained usingour PCR technique to detect these genes in A. hydrophilastrains isolated from diseased sh and from the aqua-culture environment. Neither unspecic products norpositive reactions to the A. hydrophila strain ATCC 7966Twere found with this methodology. Previous studies haveused hybridization procedures to identify the T3SS inhuman and sh isolates of Aeromonas spp. (Burr et al.,

aquaculture environments, as detected by PCR.

Total (%)

G H I J K L

13 6 40 (62.5)

0 0 4 (6.25)

6 2 20 (31.25)

2 1 1 1 13 (32.5)

0 0 0 0 0 (0)

1 0 2 3 27 (67,5)

19 3 1 3 4 8 104

ignicantly higher than that from strains from aquaculture environments

(P = 0.243).2002; Stuber et al., 2003; Chacon et al., 2004). However,these methods are time consuming, labor intensive, andexpensive. In contrast, the PCR reaction developed hereinwas fast, highly specic, had high reproducibility, and waseasily implemented in the laboratory. Thus, our techniqueis a feasible method to detect T3SS in A. hydrophila.

Previous studies showed that the T3SS is related tobacterial pathogenesis and its presence can be used as anindicator of virulence (Stuber et al., 2003; Chacon et al.,2004). Our results demonstrated that the presence of theT3SS is widespread in A. hydrophila, since T3SS genes weredetected in all farms independent of the strain origin (i.e.,clinical cases or the environment). Furthermore, the higherfrequency of the ascV+/aopB+ prole observed in strainsfrom diseased sh suggests that this system can increasethe bacterial infectivity. Similar results of T3SS frequencywere obtained by Vilches et al. (2004), who evaluated

differ

erime

pB, (

G.A. Carvalho-Castro et al. / Veterinary Microbiology 144 (2010) 371376 375mesophilic Aeromonas isolated from environment andclinical cases of human diseases. The prole ascV+/aopB+ insome environmental strains could be explained due toubiquitous condition of A. hydrophila in aquatic environ-ments associated with the abilities of bacteria to transferhorizontally the T3SS (Troisfontaines and Cornelis, 2005).The relationship between the presence of the T3SS inisolates from different sources and A. hydrophila virulenceto sh is unclear.

Table 3

Results of experimental infection with A. hydrophila strains belonging to

Straina Source Typeb ascV/aopB

AE 049-02 Water supply +/+

AE 067-02 Kidney +/+

AE 110-02 Kidney +/+

AE 111-02 Kidney +/+

AE 152-02 Wounded n +/+

AE 288-03 Kidney +/+

AE 344-03 Pond water +/+

AE 403-04 Brain +/+

AE 406-04 Kidney +/+

AE 080-02 Pond water /AE 190-02 Spleen /AE 220-02 Water supply /AE 221-02 Water supply /AE 225-02 Water supply /AE 266-03 Water supply /AE 333-03 Pond water /AE 410-04 Kidney /AE 413-04 Kidney /AE 178-02 Kidney /AE 179-02 Kidney +/AE 203-02 Liver +/AE 204-02 Liver +/a Strains randomly chosen according to ascV/aopB gene proles for expb (+/+) positive to both genes, (+/) positive to ascV and negative to aoc Main clinical signs observed in challenged groups.In this work, four ascV+/aopB strains were found,which suggests that the aopB genewas absent or that theseisolates have mutations in this gene and thus could not bedetected by PCR. There are no previous descriptions aboutthis phenomenon in A. hydrophila isolates. Even though,the occurrence of a degenerate T3SS that become thevirulence factor afunctional have been reported inEscherichia coli, seeming to be a common situation inGram negative bacteria (Tobe et al., 2006).

Based on themortality rates and clinical signs observed,ascV+/aopB+ strains were more virulent than ascV+/aopBand ascV/aopB strains in the experimental infection.Although strains belonged in the three genetic prolescaused disease, different clinical signs and illness evolutionwere observed among them. The isolates ascV+/aopB+promoted sudden disease, with fast evolution, character-ized bywidespread hemorrhage on the body surface and inthe mucosa. In contrast, sh challenged with ascV/aopBor ascV+/aopB isolates showed only apathy, anorexia, andskin darkening accompanied by low mortality rates. Thepathogenesis of A. hydrophila is multifactorial, and pre-vious studies have demonstrated the involvement of manyvirulence factors including other secretion systems as T6SS(Wong et al., 1998; Yu et al., 2005; Suarez et al., 2008).Nevertheless, the ascV+/aopB strains seemed to be poorlyadapted to promoting sh infection than the ascV+/aopB+strains. These results agree with those from studies of A.hydrophila T3SS gene mutants that showed a decrease invirulence and lower cytotoxicity (Vilches et al., 2004; Yuet al., 2004; Sha et al., 2005). The low mortality ratesobserved in challenge assays performed with four ascV+/aopB+ strains (Table 3) might be associated with possiblevariations in the secretion mechanism of the effectortoxins or with different behaviors in global gene expres-sion, compared to other ascV+/aopB+ strains that induced

ent genetic proles of ascV and aopB genes.

Main clinical sign observedc Mortality%

Cutaneous hemorrhage 12.5

Cutaneous hemorrhage 75

Cutaneous hemorrhage 75

Erratic swimming and exophthalmia 87.5

Cutaneous hemorrhage 25

Erratic swimming 12.5

Cutaneous hemorrhage 25

Cutaneous hemorrhage 25

Cutaneous hemorrhage 87.5

0

Anorexia, lethargy, darkening skin 37.5

Anorexia and lethargy 12.5

0

0

Anorexia, lethargy, darkening skin 12.5

0

Anorexia, lethargy, darkening skin 12.5

0

0

0

0

Anorexia and lethargy 12.5

ntal infection.

/) negative to both genes.high mortality (Francis et al., 2002; Meja et al., 2008).Overall, the results obtained from the experimentalinfections suggest that A. hydrophila, similar of veriedin E. coli, can show different subpopulations or pathotypeswith variable ability to infect the hosts and induce clinicalsigns (Kaper et al., 2004).

We conclude that the screened genes were morefrequent in A. hydrophila strains associated with clinicaldisease in sh than in the environment and that ascV+/aopB+ strains were more virulent to Nile tilapia comparedto the other two genetic proles. The PCR protocolsdeveloped for both genes proved to be good assays for T3SSscreening and could be used in routine analyses for A.hydrophila virulence typing.

Competing interest

The authors declare no competing nancial interests.

Acknowledgments

This work was supported by grants FAPEMIG CVZ APQ01734/08, CNPq (INCT 573899/2008-8). We would like tothank FAPEMIG and CAPES for the student fellowships. Wealso thank Dirceia A. da Costa Custodio for her technicalassistance.

Appendix A. Supplementary data

Supplementary data associated with this article can be

found, in the online version, at doi:10.1016/j.vetmic.

2010.01.021.

References

Abbott, S.L., Cheung, W.K.W., Janda, J.M., 2003. The genus Aeromonas:biochemical characteristics, atypical reactions and phenotypic iden-tication schemes. J. Clin. Microbiol. 41, 23482357.

Borrel, N., Acinas, S.G., Figueras, M.J., Martinez-Murcia, A.J., 1997. Identi-cation of Aeromonas clinical isolates by restriction fragment lengthpolymorphism of PCR-amplied 16S rRNA genes. J. Clin. Microbiol. 35(7), 16711674.

Burr, S.E., Stuber, K., Whali, T., Frey, J., 2002. Evidence for a type IIIsecretion system in Aeromonas salmonicida subsp salmonicida. J. Bac-teriol. 151, 21112118.

Chacon, M.R., Soler, L., Groisman, E.A., Guarro, J., Figueras, M., 2004. TypeIII secretion system genes in clinical Aeromonas isolates. J. Clin.Microbiol. 42 (3), 12851287.

Cornelis, G.R., 2006. The type III secretion injectisome. Nat. Rev.Microbiol.4, 811825.

Edberg, S.C., Browne, F.A., Allen, M.J., 2007. Issues for microbial regula-tion: Aeromonas as a model. Crit. Rev. Microbiol. 33, 89100.

Fang, H.M., Ge, H., Sin, Y.M., 2004. Cloning, characterization and expres-sion of Aeromonas hydrophilamajor adhesion. Fish Shellsh Immunol.16, 645658.

Francis, S.M., Wolf-Watz, H., Forsberg, A., 2002. Regulation of type IIIsecretion systems. Curr. Opin. Microbiol. 5, 166172.

Kaper, J.B., Nataro, J.P., Mobley, H.L., 2004. Pathogenic Escherichia coli. Nat.Rev. Microbiol. 2, 123140.

Meja, E., Bliska, J.B., Viboud, G.I., 2008. Yersinia controls type III effectordelivery into host cells by modulating rho activity. PLoS Pathog. 4 (1),e3 00140024.

Mueller, C.A., Broz, P., Cornelis, G.R., 2008. The type III secretion system tipcomplex and translocon. Mol. Microbiol. 68 (5), 10851095.

Seshadri, R., Joseph, S.W., Chopra, A.K., Sha, J., Shaw, J., Graf, J., et al., 2006.Genome sequence of Aeromonas hydrophila ATCC 7966TT: Jack of alltrades. J. Bacteriol. 188, 82728282.

Sha, J., Pillai, L., Fadl, A.A., Galindo, C.L., Erova, T.E., Chopra, A.K., 2005. Thetype III secretion system and cytotoxic enterotoxin alter the virulenceof A. hydrophila. Infect. Immun. 73 (10), 64466457.

Sierra, J.C., Suarez, G., Sha, J., Foltz, S.M., Popov, V.L., Galindo, C.L., et al.,2007. Biological characterization of a new type III secretion systemeffector from a clinical isolate of Aeromonas hydrophilaPart II.Microb. Pathog. 43, 147160.

Suarez, G., Sierra, J.C., Sha, J., Wang, S., Erova, T.E., Fadl, A.A., et al., 2008.Molecular characterization of a functional type IV secretion systemeffector from a clinical isolate of Aeromonas hydrophila. Microb.Pathog. 44 (4), 344361.

Stuber, K., Frey, J., Kuhnert, P., 2003. Detection of type III secretion genesas a general indicator of bacterial virulence. Mol. Cell. Probes 17, 2532.

Tobe, T., Beatson, S.A., Taniguchi, H., Abe, H., Bailey, C.M., Fivian, A., et al.,2006. An extensive repertoire of type III secretion effectors in Escher-ichia coliO157 and the role of lambdoid phages in their dissemination.Proc. Natl. Acad. Sci. U.S.A. 103, 1494114946.

Troisfontaines, P., Cornelis, G.R., 2005. Type III secretion: more systemsthan you think. Physiology 20, 326339.

Vilches, S., Urgell, C., Merino, S., Chacon,M.R., Soler, L., Castro-Escarpulli, G.,2004. Complete type III secretion system of a mesophilic Aeromonashydrophila strain. Appl. Environ. Microbiol. 70 (11), 19161919.

Vivas, J., Carracedo, B., Riano, J., Razqin, B.E., Lopez-Fierro, P., Acosta, F.,Naharro, G., Villena, A.J., 2004. Behavior of an Aeromonas hydrophilaaroA live vaccine in water microcosms. Appl. Environ. Microbiol. 70,27022708.

Wong, C.Y., Heuzenroeder, M.W., Flower, R.L., 1998. Inactivation of twohaemolytic toxin genes in Aeromonas hydrophila attenuates virulencein a suckling mouse model. Microbiology 144, 291298.

Yu, H.B., Sirinivasa Rao, P.S., Lee, A.C., Vilches, S., Merino, S., Tomas, J.M.,Leung, K.Y., 2004. A type III secretion system is required for Aeromonashydrophila AH-1 pathogenesis. Infect. Immun. 72 (3), 12481256.

Yu, H.B., Zhang, Y.L., Lau, Y.L., Yao, F., Vilches, S., Merino, S., et al., 2005.Identication and characterization of putative virulence genes andgene clusters in A. hydrophila PPD134/91. Appl. Environ. Microbiol. 71

G.A. Carvalho-Castro et al. / Veterinary Microbiology 144 (2010) 371376376(8), 44694477.

Detection of type III secretion system genes in Aeromonas hydrophila and their relationship with virulence in Nile tilapiaIntroductionMaterials and methodsBacterial strainsDNA extractionOligonucleotide primersPCR amplificationDNA sequencingExperimental infectionStatistical analysis

ResultsPCR standardizationFrequency of ascV and aopB genesExperimental infection

DiscussionCompeting interestAcknowledgmentsSupplementary dataReferences